Streak Recognition for a Three-Dimensional Volumetric Particle Tracking Velocimetry System

摘要

A three-dimensional volumetric particle tracking veloci-metry (VPTV) system is introduced for characterizing the full-scale velocity field of a three-dimensional domain. A streak recognition algorithm has been developed for this VPTV system to automatically detect the particle streaks in the image coordinate system with the extraction of their complete parameters, such as length, location, and angle. More than 90% of seeding particle streaks are identified and located with the coordinates of their end points in the image. A powerful pattern recognition technique—Hough transform—is used to develop a unique crossover streak recognition algorithm. This streak recognition algorithm increases the image processing efficiency by avoiding visual inspection of the crossover streaks. It makes characterization of complicated flow fields, involving large amounts of image processing by the VPTV system, feasible. After integrating the streak recognition into the VPTV system, a full-scale, five-row section mockup of a Boeing 767-300 aircraft cabin is measured.